Apple Releases ProRes QuickTime Decoder 1.0

[LethalWolfe]

Just as Analog kid and others pointed out, it most definitely IS a weasel word. "Lossless" should NEVER be used when a codec is not mathematically lossless. Just because I could rip a CD track into a 1024kbps AAC file and not be able to ever tell the difference even If I encode it a 1000 times doesn't mean I can say it "sounds lossless". If it throws away any amount of information, then it's a LOSSY codec.

Why can't you say it 'sounds lossless'? 'Sounding lossless' and 'being lossless' aren't the same thing but they both still exist. Thanks to improving encoding technology and CPU power we have robust, compressions that while mathematically lossy they are able to retain their quality similar to uncompressed codecs for practical purposes. So why not make a distinction for this kind of codec? There is uncompressed, lossy that looks and acts like uncompressed in all but extreme uses, and lossy that looks and acts lossy. You say weasel word, I say a distinction that makes my life a little easier when trying to develop post production workflows.

And if we really want to get picky every recorded image is compressed from the start because nothing can record an infinite amount of information.


Lethal

 

[Tosser]

Why can't you say it 'sounds lossless'? 'Sounding lossless' and 'being lossless' aren't the same thing but they both still exist. Thanks to improving encoding technology and CPU power we have robust, compressions that while mathematically lossy they are able to retain their quality similar to uncompressed codecs for practical purposes. So why not make a distinction for this kind of codec? There is uncompressed, lossy that looks and acts like uncompressed in all but extreme uses, and lossy that looks and acts lossy. You say weasel word, I say a distinction that makes my life a little easier when trying to develop post production workflows.

I have a better idea, since it's lossy, let's call it lossy. If you want to differentiate between low bitrate mp3 and so forth, let's call this a high bitrate lossy format, because that is what it is. No more, no less.

Since it's not lossless, there's no need to use the term "lossless" as a marketing term, watering it down with preemptions, only to polute the waters.

And if we really want to get picky every recorded image is compressed from the start because nothing can record an infinite amount of information.

I don't think you know what compressed means.

 

[Analog Kid]

...
Thanks to improving encoding technology and CPU power we have robust, compressions that while mathematically lossy they are able to retain their quality similar to uncompressed codecs for practical purposes. So why not make a distinction for this kind of codec? There is uncompressed, lossy that looks and acts like uncompressed in all but extreme uses, and lossy that looks and acts lossy.
...

Distinctions are made where they make sense. In this case, the distinction is made between codecs that can be reversed without loss of information, and those that can't. Those that can are called lossless, and those that can't are called lossy.

What you are describing isn't a distinction, it's a continuum. There is more or less similar to uncompressed, there are different quality metrics for different uses. Breaking codecs into categories based on continuous parameters such as these require an arbitrary threshold rather than being inherent in the codec itself.

ProRes is a lossy codec suitable for *insert use*. If the industry wants to attach a word to codecs that a trained professional can't distinguish from raw on a given display after x encoding generations, that's fine too. Just don't call it lossless if it isn't.

And if we really want to get picky every recorded image is compressed from the start because nothing can record an infinite amount of information.

You're confusing image capture with image compression, and to get even more pedantic you're confusing compression with transformation.

Image capture will almost always be lossy for the reasons you state.

 

[yoak]

Nitpicking (is that even a word?) aside, this is great news to me/us as we have one MP, but my partner uses Premiere on a PC. Finally we can exchange files in an easy way (although I won´t belive it until I see it)
There always seems to be a glitch somewhere in a process like this, but I´m hoping and I´m quite exited.
It will make life much easier and hopefully free up some of the time we spend on moving footage back and forth

 

[Virtuoso]

this is great news to me/us as we have one MP, but my partner uses Premiere on a PC. Finally we can exchange files in an easy way

There is a catch - if you make any changes in Premiere, you won't be able to stay using ProRes. The codec is only a decoder, not an encoder, so whilst you can read the files on a PC, you will need to use another codec (or uncompressed) if you send the files back to the Mac.

 

[winterspan]

I have a better idea, since it's lossy, let's call it lossy. If you want to differentiate between low bitrate mp3 and so forth, let's call this a high bitrate lossy format, because that is what it is. No more, no less. *snip*

Distinctions are made where they make sense. In this case, the distinction is made between codecs that can be reversed without loss of information, and those that can't. Those that can are called lossless, and those that can't are called lossy.

What you are describing isn't a distinction, it's a continuum. There is more or less similar to uncompressed, there are different quality metrics for different uses. Breaking codecs into categories based on continuous parameters such as these require an arbitrary threshold rather than being inherent in the codec itself. *snip*

Both great posts. Perhaps someone needs to come up with a more suitable term for the industry to use, and sooner rather than later lest we start seeing every company using some inane variant of "visually lossless".

 

[yoak]

There is a catch - if you make any changes in Premiere, you won't be able to stay using ProRes. The codec is only a decoder, not an encoder, so whilst you can read the files on a PC, you will need to use another codec (or uncompressed) if you send the files back to the Mac.

Guess I should have read more about it, but I was too excited.
Still have to figure out workarounds then.

 

[netdog]

And if we really want to get picky every recorded image is compressed from the start because nothing can record an infinite amount of information.

Lethal

Lossless is generally applied to the acquisition and compression of digital sources and means that from the compressed file, an original can be recreated that matches that file bit for bit. Nobody talks about Lossless archiving of LPs or analog photographs.

As for DVD and CD conversion, yes, lossless means lossless in regards to the original source, though indeed films are compressed to make DVDs (substantially compressed in fact) and music played in a studio is dramatically compressed as well, starting with the microphones or other devices used to capture sound.

 

[ThePhage]

I'm not a compression expert, but I am a video editor and a phrase such as "visually lossless" speaks to me (it has since I read a description of Cineform's codec 2 years ago). It doesn't mislead me. I immediately understand that it refers to a compressed format, otherwise it would be termed "Uncompressed." I know that this compression scheme will be throwing away information, but the "visual" prefix is an indicator of how important that information is (or isn't).

Today lossless compression isn't used for video in any serious degree. Video is commonly either uncompressed or compressed in a lossy fashion. Uncompressed video requires large amount of storage capacity (especially at HD resolution, high sample depth, and with no color subsampling). So the question becomes, does a lossy compression scheme hold enough visual quality after x number of render generations before losing the benefit of smaller file sizes.

However, I can understand how audiophiles who are used to slinging around the "lossless/lossy" terms in endless online debates could be misled. Perhaps the marketing people decided that "Almost Lossless" didn't quit fit the bill.

 

[Tosser]

I'm not a compression expert, but I am a video editor and a phrase such as "visually lossless" speaks to me (it has since I read a description of Cineform's codec 2 years ago). It doesn't mislead me. I immediately understand that it refers to a compressed format, otherwise it would be termed "Uncompressed." I know that this compression scheme will be throwing away information, but the "visual" prefix is an indicator of how important that information is (or isn't).

Funny. On one hand you agree that it isn't lossless, and that you don't get the "lossless" part from the monicker. Then why on earth do you think that "lossless" is even a part of the name?
You also agree that this is a lossy format, so why on earth not call it what it is?
The reason is of course, that what you claim - or infer, if you will - isn't true as a whole. It IS a marketing term to make it sound like it is something it isn't all the while covering their respective arses.

Today lossless compression isn't used for video in any serious degree. Video is commonly either uncompressed or compressed in a lossy fashion. Uncompressed video requires large amount of storage capacity (especially at HD resolution, high sample depth, and with no color subsampling). So the question becomes, does a lossy compression scheme hold enough visual quality after x number of render generations before losing the benefit of smaller file sizes.

As mentioned, it IS a lossy format (YOU even mention it), and thus the only reason for including the term "lossless" is to make belief it is something it isn't.

However, I can understand how audiophiles who are used to slinging around the "lossless/lossy" terms in endless online debates could be misled. Perhaps the marketing people decided that "Almost Lossless" didn't quit fit the bill.

I, for one, am not an audiophile. Your put-down really is misplaced and ignorant.

I get that you don't care either way, but I really don't see the need to muddy the waters by introducing a lossy format, but calling it "[virtually] lossless". It's a high bit rate lossy format. That's it.
On principle, I wouldn't like cameras which crops to be called "visually full frame" either. It 's a matter of using the correct terms. Oh, and in short - with regards to the last paragraph of yours: Don't be daft.

 

[jimis83]

greek subtitles

hello i have a mac and i have movies that support subtitles that you can download from different websites. but when i open the movie and watch they appear like chinese/greek.if i choose english subs all is ok.what is wrong??i use the vlc player!!!!

 

[LethalWolfe]

I have a better idea, since it's lossy, let's call it lossy. If you want to differentiate between low bitrate mp3 and so forth, let's call this a high bitrate lossy format, because that is what it is. No more, no less.

But when the next logical question is "How well does the image compare to the original?" and the answer is "visually there is no loss of image info" why give these kinds of codecs an obtuse, generic moniker? For practical purposes, IMO, "visually lossless" is a better name than "high bit rate lossy format" because "visually lossless" tells me two things. First, it tells me that the codec is lossy. Second, it tells me that the codec, even though lossy, shouldn't degrade the visual quality of the image compared to an uncompressed codec under typical use. There is a difference between visually lossy and mathematically lossy so I don't see the problem using a term that takes that difference into account.

Some posters are saying they've never heard of "visually lossless" before the start of this thread and think it is an inaccurate, marketing term. I'm saying I've seen the term for around two years now and I find it a helpful term to use when discussing codecs. To each their own. In the end though I'm more interested in what the image really looks like than what an engineer in a lab w/a calculator says it should look like. For example, I know some guys that go HDV->DVCPro HD->HDCAM SR (for final delivery to the network) because their shop is built around a DVCPro HD workflow. On paper this doesn't look like a good workflow. In practice though it works w/o a hitch for these guys as they've never had a show fail quality control due to that additional transcode. I attempt to stay on top of the general aspects of codecs so I can remain aware of what's happening to the footage as it moves along from production to post to final delivery, but I don't let semantics get in the way of a viable workflow. The proof is in the pudding and if the final product looks solid I consider it a good workflow even if 'on paper' it doesn't look like the best way to do things.

The reason I brought up compression during acquisitions was just an attempt to illustrate that in practice both "uncompressed" and "compressed" are moving targets in terms of quality. For example, given comparable cameras a compressed DV signal is on par w/an 'uncompressed' signal captured from a BetaSP tape (Link-scroll down slightly). As technology keeps getting bigger, faster, and cheaper the traditional off-line (lossy/compressed) and on-line (uncompressed) workflow, necessitated by technological limitations, will keep getting blurred and 'compressed' isn't the four letter word it used to be.


Lethal